Skip to content Skip to navigation

Session 6: Psychology and Economics

August 14-16, 2019 | Koret-Taube Conference Center, 366 Galvez St, Stanford

The purpose of this workshop is to bring together researchers working on issues at the intersection of psychology and economics. The workshop focuses on evidence of and explanations for non-standard choice patterns, as well as the positive and normative implications of those patterns in important economic decision-making contexts such as lifecycle consumption and savings, labor supply, effort provision, financial contracting, and firm pricing. Many past presenters at this workshop have built upon insights from other disciplines, including psychology and sociology. Theoretical, empirical, and experimental studies will be included.

Organizers: B. Douglas Bernheim (Stanford), John Beshears (Harvard Business School), Vincent Crawford (Oxford), David Laibson (Harvard) and Ulrike Malmendier (University of California, Berkeley)

In this Session

Aug 14 | 9:00 am to 9:30 am

Check-in | Breakfast

Aug 14 | 9:30 am to 10:05 am

Default Options and Retirement Saving Dynamics

Presented by: Taha Chouckhmane (Yale University)

Does automatic enrollment in retirement savings plans increase lifetime wealth accumulation and welfare? I document that employees offset the short-run positive effect of automatic enrollment by saving less in the future. Consequently, a lifecycle model estimated on data from 34 U.S. 401(k) plans predicts that the long-term effect of auto-enrollment on wealth is negligible except at the bottom of the lifetime earnings distribution. The model’s long-run predictions are supported by its out-of sample performance in predicting behavior in 86 other 401(k) plans and nationally representative U.K. data. The observed inertia at the savings default is explained by an opt-out cost of around $250. My estimate is smaller than the thousands of dollars estimated in previous studies because, in my fully dynamic model, non-autoenrolled workers can compensate for not contributing now by contributing more later. Automatic enrollment improves social welfare if the policymaker is more patient than individuals or puts more weight on low-income individuals’ utility.

Aug 14 | 10:05 am to 10:40 am

The Retirement-Consumption Puzzle: New Evidence from Personal Finances

Presented by: Michaela Pagel (Columbia GSB)
Co-Author(s): Arna Olafsson (Copenhagen Business School)

This paper uses transaction-level data on individual spending, income, account balances, and credit limits from a personal finance management software provider to investigate how expenditures, liquid savings, and consumer debt change around retirement. The longitudinal nature of our data allows us to include individual fixed effects and identify the effect of individuals transitioning into retirement. We provide new evidence on the
retirement-consumption puzzle and on whether individuals save adequately for retirement. We find that, upon retirement, individuals reduce their spending in both work-related and leisure categories. However, even for the most detailed categories, we cannot tell conclusively which expenses are definitely work related or not. We thus look at household finances and find that individuals delever upon retirement by reducing consumer debt and increasing liquid savings. We argue that these findings are difficult to rationalize via, for example, work-related expenses. An agent with standard preferences would save before retirement because of the expected fall in income, and dissave after retirement, rather than the exact opposite. Additionally, we document the same patterns in commonly used US survey datasets and generate the joint observation of decreasing income and consumption but increasing savings upon retirement in a life-cycle model with non-standard preferences.

Aug 14 | 10:40 am to 11:15 am


Aug 14 | 11:15 am to 11:50 am

Does Financial Strain Lower Worker Productivity?

Presented by: Frank Schilbach (MIT)
Co-Author(s): Supreet Kaur (UC Berkeley), Sendhil Mullainathan (University of Chicago) and Suanna Oh (Columbia University)

This paper empirically tests for a direct causal impact of financial strain on worker productivity. We randomize the timing of income receipt among Indian workers who earn piece rates for manufacturing tasks: some workers receive their wages on earlier dates, altering when cash constraints are eased while holding overall wealth constant. Workers increase productivity by 5.3% on average in the days after cash receipt. The impacts are concentrated among poorer workers in the sample, who increase output by over 10%. This effect of cash on hand on productivity is not explained by mechanisms such as gift exchange, trust in the employer, or nutrition. We present positive evidence that productivity increases are mediated through lower attentional errors in production, indicating a role for improved cognition after cash receipt. Finally, directing workers’ attention to their finances via a salience intervention produced mixed results. Such primes do not appear to work effectively as a way to test the impact of worries as they also induce effort responses. Taken together, our results indicate a direct relationship between financial constraints and worker productivity and suggest that psychological channels mediated through attention play a role in this relationship.

Aug 14 | 11:50 am to 12:25 pm

Choice-Set-Dependent Preferences and Salience: An Experimental Test

Presented by: Jason Somerville (Cornell University)

This paper investigates whether the composition of the choice set alters the relative importance of quality and price. I derive and test predictions that emerge under models of choice-setdependent preferences which posit that attributes may be overweighted due to specific features of the choice set. In a laboratory experiment, I vary the prices of two vertically differentiated products to flexibly estimate attribute weights. I document large and systematic shifts in demand away from the surplus-maximizing option consistent with the relative thinking model of Bushong, Rabin, and Schwartzstein (2017). Structural estimates imply substantial heterogeneity in the magnitude but not the direction of choice-set distortions.

Aug 14 | 12:25 pm to 1:45 pm


Aug 14 | 1:45 pm to 2:20 pm

The Role of Memory in Belief Formation

Presented by: Tanya Rosenblat (University of Michigan)
Co-Author(s): Markus Mobius (MSR) and Pierre-Luc Vautrey (MIT)

A growing body of research has documented both systematic and motivated belief biases in decision making such as correlation neglect and wishful thinking. A better understanding of how people aggregate signals and form beliefs in the first place is crucial for understanding the origin of these biases and potential policy interventions. In this paper, we explore the role of memory in belief formation. We design a simple experiment where participants read a set of news sources which report naturalistic positive and negative facts about an artificial company. We provide enough information to participants about the fact-generating process so that they can form a Bayesian posterior on whether the company is "good" or "bad". We then (a) elicit participants' beliefs about the quality of company based on these narratives and (b) present participants with a superset of facts to determine whether they recognize previously seen facts. Just like real newspapers, facts can be repeated across several news sources. Unlike most of the existing literature on correlation neglect, our experiment introduces repetition through actually repeated signals which we argue is more natural than merely correlated signals. Our design therefore allows us to explore the role of recognition and recall in belief formation which are two fundamental concepts in cognitive psychology. We find that participants exhibit good recognition which is only slightly increasing by repetition of the signal. Participants slightly overweigh repeated signals (beyond what is accounted for by the greater recognition) but do not fully double-count. We propose two models of belief formation where agents either "replay" recalled signals or keep track of summary counts using the recognition channel and discuss how our experiment can distinguish between these models. Our framework can also help with understanding how individual (singleagent) learning differs from social learning.

Aug 14 | 2:20 pm to 2:55 pm

The “Fake News” Effect: An Experiment on Motivated Reasoning and Trust in News

Presented by: Michael Thaler (Harvard University)

When people receive information about controversial issues such as immigration, upward mobility, and racial discrimination, the information evokes both what they currently believe and what they are motivated to believe. This paper theoretically and experimentally explores the impact of motivated beliefs on inference: motivated reasoning. I analyze a model of motivated reasoning in which people misupdate from information by treating their motivated beliefs as an extra signal. To test predictions of the model, I create a new experimental design in which people make inferences about the veracity of customized news sources. This design constructs an environment in which subjects’ subjective likelihoods and priors are fixed but motivated beliefs exogenously vary; motivated reasoning is therefore identified from Bayesian updating, as well as from prior- and likelihood-misweighting biases. Using a large online experiment, I find evidence for politically-driven motivated reasoning on eight topics: immigration, upward mobility, racial discrimination, crime, gender and math ability, climate change, gun laws, and beliefs about other subjects. I also find ego-driven motivated reasoning on beliefs about own performance, but only for men. I additionally show how motivated reasoning leads people to become more polarized, less accurate, and overconfident in their beliefs.

Aug 14 | 2:55 pm to 3:30 pm


Aug 14 | 3:30 pm to 4:05 pm

Frequent Price Changes, Perceived Inflation, and Inflation Experiences

Presented by: Francesco D’Acunto (University of Maryland)
Co-Author(s): Ulrike Malmendier (UC Berkeley), Juan Ospina (University of Chicago) and Michael Weber (University of Chicago)

How do households form their in inflation expectations? We show the price changes individual consumers observe while shopping are a key determinant of their expectation for overall inflation. We use individual non-durable consumption choices in the Nielsen Homescan Panel to construct household-level measures of perceived inflation. We find perceived price changes help explain inflation expectations across individuals and within individuals over time. The frequency of purchases not the expenditure share determines the importance of perceived price changes for inflation expectations. The effect is stronger for individuals that shop less frequently, and hence are more likely exposed to several and larger price changes in their typical shopping trip. The effect is also stronger for individuals whose uncertainty about inflation is higher and who self report to not rely on the media when forming expectations. Because individual inflation expectations shape economic decisions, central banks' focus on core inflation instead of the heterogeneous price changes in households' non-durable bundles might lead to systematic policy mistakes.

Aug 14 | 4:05 pm to 4:40 pm

The Biases of Others

Presented by: Kristof Madarasz (LSE)
Co-Author(s): David Danz (University of Pittsburg) and Stephanie Wang (University of Pittsburgh)

When social beliefs are systematically biased, understanding how people think about the biases of others is essential since biased beliefs about the beliefs of others and the anticipation of the same biased tendency in the thinking of others are directly linked. We find that while people fail to engage in sufficient mental perspective taking and naively project their information onto differentially-informed others, they also anticipate that differentiallyinformed others mistakenly project onto them. In particular, we test a tight one-to-one relationship between the extent to which the typical person projects onto others, ρ, and the partial extent to which she anticipates but underestimates the projection of others onto her, ρ2, as implied by projection equilibrium. We find that most people are partially biased and partially anticipate the biases of others and that the structure of this psychological phenomena is remarkably consistent with the proposed portable and tight idea of social beliefs arising from a coherent but fully egocentric belief hierarchy with a partial probabilistic adjustment to the truth. Applications to defensive agency and law and economics are discussed.

Aug 14 | 6:30 pm to 9:30 pm


Aug 15 | 9:00 am to 9:30 am


Aug 15 | 9:30 am to 10:05 am

The Dynamics of Norm Formation and Norm Decay

Presented by: Ernst Fehr (University of Zurich)
Aug 15 | 10:05 am to 10:40 am


Presented by: Tore Ellingsen (Stockholm School of Economics)
Co-Author(s): Erik Mohlin (Lund University)

We develop a formal theory of decency. Shared values and understandings give rise to social norms. Norms may mandate collectively optimal behavior, but they need not do so. Furthermore, behavior can be affected by social values even if it stops short of norm compliance. Seeking stronger predictions, we propose a structural model of social values; society endorses efficiency and equality, but condemns ill-gotten gains. The model implies that decent people will tend to avoid situations that encourage prosocial behavior. It also rationalizes the existence of willful ignorance, intention-based negative reciprocity, and betrayal aversion.

Aug 15 | 10:40 am to 11:15 am


Aug 15 | 11:15 am to 11:50 am

The (In)Elasticity of Moral Ignorance

Presented by: Marta Serra-Garcia (UC San Diego)
Co-Author(s): Nora Szech (Karlsruhe Institute of Technology)

Ignorance enables individuals to act immorally. This is well known in policy circles, where there is keen interest in lowering moral ignorance. In this paper, we demonstrate the relevance of demand elasticity to ignorance
by showing that small monetary incentives can significantly reduce ignorance. We contrast monetary incentives with social norms, which have little impact on ignorance and actually increase ignorance in less moral individuals. Using a longitudinal design, we document that ignorance is persistent across moral contexts and through time. We propose and structurally estimate a simple behavioral model in line with our findings.

Aug 15 | 11:50 am to 12:25 pm

Chosen Preferences

Presented by: Doug Bernheim (Stanford)
Aug 15 | 12:25 pm to 1:45 pm


Aug 15 | 1:45 pm to 2:20 pm

Exaggerating to Break-Even: Reference-Dependent Moral Hazard in Auto Insurance Claims

Presented by: Jaimie Lien (The Chinese University of Hong Kong)
Co-Author(s): Feng Gao (Tsinghua University) and Jie Zheng (Tsinghua University)

The effects of asymmetric information are often difficult to detect empirically, such as in insurance settings (Chiappori, Jullien, Salanie and Salanie, 2006). We show that allowing for the possibility of reference dependent preferences can assist with this challenge. Using detailed auto insurance claims data and adopting the methodology of Allen, DeChow, Pope and Wu (2017) which studies reference-dependence in marathon finishing times, we show that policyholders exaggerate their damage claims in a manner consistent with referencedependent moral hazard. Consistent with our theoretical model of reference-dependent claims, we find stronger reference-dependent behavior among policyholders with higher premium levels and low risk policyholders. Also consistent with the model, policyholders are more reference-dependent in their first claims than in subsequent claims, and in the second half of the coverage year compared to the first half. The detected reference point is the amount of the original premium paid - in other words, in claiming their accident damages, policyholders seek to claim back the original price of their insurance. We do not find robust patterns in the concentration of insurance claims at round number percentages greater than or less than the full premium amount, such as 90% or 110%, nor at round number nominal amounts. This implies that the unnatural claim behavior is motivated significantly by loss aversion regarding the premium paid rather than merely wanting to increase one’s claims

Aug 15 | 2:20 pm to 2:55 pm

Meaningful Theorems: Nonparametric Analysis of Reference-Dependent Preferences

Presented by: Vince Crawford (Oxford)
Co-Author(s): Laura Blow (University of Surrey) and Ian Crawford (University of Oxford)
Aug 15 | 2:55 pm to 3:30 pm


Aug 15 | 3:30 pm to 4:05 pm

Eliminating Equilibrium Pathologies in Models with Present-Biased Discounting: The β-δ-Δ Sweet Spot

Presented by: Peter Maxted (Harvard)
Aug 15 | 4:05 pm to 4:40 pm

Payday Lending, Naïve Present Focus, and Consumer Protection

Presented by: Hunt Allcott (New York University)
Aug 15 | 6:30 pm to 9:30 pm


Aug 16 | 9:00 am to 9:30 am


Aug 16 | 9:30 am to 10:05 am

Estimating Biases in Smoking Cessation: Evidence from a Field Experiment

Presented by: Matthew Levy (London School of Economics)
Co-Author(s): Frank Chaloupka (University of Illinois at Chicago) and Justin S. White (University of California, San Francisco)

We undertook a randomized field experiment to quantify three behavioral biases that may affect consumers of addictive goods: present-biased preferences, naïve beliefs regarding present bias, and projection-biased beliefs over future abstinence. These biases compose a minimal set of departures from the neoclassical benchmark which can accommodate both intertemporal and state-dependent mis-predictions, and have important theoretical and policy ramifications. Our experimental design employs a new lottery-based monitoring design to ensure strict incentive-compatibility at all stages, and a novel identification of subjects’ biases based on demand for partial commitment devices. We find that cigarette smokers overestimated their likelihood of future abstinence by more than 100%, consistent with partially-naïve present-biased preferences. In a structural analysis, we estimate present bias and beliefs about present bias to be an average = 0.62 and ˜ = 0.71, respectively, with substantial heterogeneity and a strongly positive correlation between the two at the individual level. Smokers further mispredict the effects of an abstinence intervention on their future abstinence. Inconsistent with a simple model of projection bias, our estimates imply that ex-ante smokers anticipate no effect (i.e., complete projection over long-run addiction) and ex-post believe the effect to be
marginally negative. Our estimates highlight that smokers suffer from a pernicious constellation of biases: under their own long-run preferences, smokers’ choices lead to a welfare loss of between $600 and $1200 per week.

Aug 16 | 10:05 am to 10:40 am

Incentivizing Behavioral Change: The Role of Time Preferences

Presented by: Rebecca Dizon-Ross (University of Chicago)
Aug 16 | 10:40 am to 11:15 am


Aug 16 | 11:15 am to 11:50 am

How Are Preference for Commitment Revealed?

Presented by: Dmitry Taubinsky (UC Berkeley)
Co-Author(s): Mariana Carrera (Montana State University), Heather Royer (UC Santa Barbara), Mark Stehr (Drexel University), and Justin Sydnor (University of Wisconsin at Madison)

A large literature treats take-up of commitment contracts, in the form of choice-set restrictions or penalties, as a smoking gun for (awareness of) self-control problems. This paper providesnew techniques for examining the  validity of this assumption, as well as a new approach for detecting (awareness of) self-control problems. Theoretically, we show that with some uncertainty about the future, demand for commitment contracts is closer to a special case than to a robust implication of models of limited self-control. In a field experiment with 1292 members of a fitness facility, we find that many participants take up commitment contracts both for going to the gym more and for going to the gym less, and there is a significant positive correlation in demand for these two types of contracts. This suggests that commitment contract take-up reflects, at least in part, something other than the desire to change own future behavior, such as demand effects or “noisy valuation.” Moreover, we find that commitment contract take-up is negatively related to awareness of self-control problems: a novel information treatment that increased awareness of self-control problems reduced demand for commitment contracts. We address the limitations of using commitment contracts as a measurement tool by showing that a combination of belief forecasts and willingness to pay for linear incentives provides more robust identification of limited self-control and people’s awareness of it. We use the methodology to obtain some of the first parameter estimates of partially-sophisticated quasi-hyperbolic discounting in the field.

Aug 16 | 11:50 am to 12:25 pm

Behavioral Food Subsidies

Presented by: Michael Kuhn (University of Oregon)
Co-Author(s): Brownback, Andy (University of Arkansas) and Alex Imas (Carnegie Mellon University)

We conduct a pre-registered eld experiment with low-income grocery shoppers to study the eectiveness of subsidies for encouraging healthy food purchases. Our unique design enables us to deliver subsidies and elicit choices over different types of subsidies both before and at the point of purchase. Our study examines the eect of two non-restrictive changes to the choice environment: giving shoppers greater agency over the choice of subsidies and introducing a waiting period before the shopping trip to prompt deliberation on the subsidy and food choice. Allowing shoppers to choose between healthy and less healthy subsidies does not decrease healthy food purchases compared to a restricted healthy subsidy {if anything, the effect is positive. Strikingly, we find that pairing greater agency with a waiting period nearly doubles the effectiveness of subsidies on spurring healthy food spending. This effect is large and economically meaningful given that a restricted healthy subsidy already causes a substantial increase in healthy food purchases over a control group. Exploring mechanisms, we find suggestive evidence that shoppers use the active choice of a healthy food subsidy as a costly commitment device to purchase more healthy food. Additionally, our interventions are most effective for those with more pronounced hyperbolic discounting. Given the low cost and potential scalability of our interventions, our findings have signicant implications for the design of policy tools.

Aug 16 | 12:25 pm to 1:45 pm